Add the namespace hook for GMLVQ in the model class

prototorch/models/glvq.py

@@ -191,6 +191,9 @@ class GMLVQ(GLVQ):
                                            self.hparams.latent_dim,
                                            bias=False)
 
+        # Namespace hook for the visualization callbacks to work
+        self.backbone = self.omega_layer
+
     @property
     def omega_matrix(self):
         return self.omega_layer.weight.detach().cpu()